Convergence of denoising diffusion models under the manifold hypothesis - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2022

Convergence of denoising diffusion models under the manifold hypothesis

Valentin de Bortoli
  • Fonction : Auteur
  • PersonId : 1084097
  • IdRef : 249460300

Résumé

Denoising diffusion models are a recent class of generative models exhibiting state-of-the-art performance in image and audio synthesis. Such models approximate the time-reversal of a forward noising process from a target distribution to a reference density, which is usually Gaussian. Despite their strong empirical results, the theoretical analysis of such models remains limited. In particular, all current approaches crucially assume that the target density admits a density w.r.t. the Lebesgue measure. This does not cover settings where the target distribution is supported on a lower-dimensional manifold or is given by some empirical distribution. In this paper, we bridge this gap by providing the first convergence results for diffusion models in this more general setting. In particular, we provide quantitative bounds on the Wasserstein distance of order one between the target data distribution and the generative distribution of the diffusion model.

Dates et versions

hal-03945469 , version 1 (18-01-2023)

Identifiants

Citer

Valentin de Bortoli. Convergence of denoising diffusion models under the manifold hypothesis. 2023. ⟨hal-03945469⟩
9 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More